Citation Relations are currently not too useful

[ror3d]

Is your suggestion for improvement related to a problem? Please describe. Citation relations, as currently implemented, can get citations and cited by articles for a paper, but the interface doesn't let you read the abstract to determine if the article is actually interesting, and only way one can use the information is to add it to the database, and when it's added, the new entry is set up as a Misc entry and lacks any kind of information other than the title, author and year of the entry, but no DOI or anything similar.

Describe the solution you'd like

• There should be a way to open an entry from the Citation Relations tab in the browser to be able to get more information on it without adding it to the database.
• When adding the entries from the tab to the database, more information should be included in the new entries. Ideally, any information that would be downloaded as if the entry had been added from the browser extension would be included; but at least having the DOI to be able to use the functionality from General > DOI > Get bibliographic data from DOI would be the bare minimum.

Additional context

• I have started implementing the second part, getting the DOI information into the new entry, and attempting to mark the type of entry based on the type reported by the semantic scholar API. That would solve the "bare minimum" described in the possible solution.
• I was thinking of implementing the first point by adding a button on each entry to open it in the browser for preview.
• To implement the better solution for the second point, one would need to look into how article download happens normally and trigger that when new entries are added.

Proposed Tasks

• [x] Add button to open a relation entry in the browser
• [x] Include additional information in the created entry
• [x] Way to preview some more information in the citation relations tab (already existing, just needed the extra information retrieved from SemanticScholar)
• [ ] Some more detailed way to preview the entry from the citation relations tab before importing from inside JabRef, so as not to have to open it in the browser (slow)

[Siedlerchr]

Thanks for the suggestion, that sounds really good. @JabRef/developers any ideas about the other parts?

[calixtus]

Mh, would be great if one could somehow preview the entry directly in jabref without opening a browser, because that takes time and is somewhat complicated. Maybe a tooltip or a folding option to open the entry in the list and display the abstract inside...

[ror3d]

The semantic scholar api request can include the request for the abstract, it would be a matter of showing it somehow. My suggestion was to add a button to open the DOI url mainly because that's something I believe I could do without much trouble, implementing a more complex system to show it inline is not something I'd feel comfortable attempting.

But maybe this can be done iteratively, get something working a bit better now and improve the UI when someone that knows what they are doing can put time in it.

[koppor]

Thank you so much for raising your voice regarding that feature.

In case someone else should work on it, we need a very refined specification. Including screenshots / screen mockups. Most contributors (currently) are computer science students in their bachelors and thus having little knowledge about the domain. See for instance the comment at https://github.com/JabRef/jabref/pull/10784#issuecomment-1909135492. -- We core developers are (unfortunately) very busy answering and categorizing issues, as well as finalizing incoming pull requests.

[koppor]

Related: https://github.com/JabRef/jabref/issues/10414

[ror3d]

PR #11016 addresses several of the points raised here. As can be seen below, the entries in the references list now also include the abstract, as proposed by calixtus, and a button to open the paper's website in the browser:

I'm unsure if there is anything else that could be done as improvements (that's why I didn't close the issue from the PR) but if it feels like this is enough then this issue can be closed.

[ThiloteE]

I haven't really followed this issue very closely. Could you please edit your initial post and mark those issues that have been implemented already and those that have not been implemented?

e.g. via - [ ] and - [x]

[ror3d]

Sure, I've added that to the issue description